home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.software-eng
- Path: sparky!uunet!cis.ohio-state.edu!zaphod.mps.ohio-state.edu!uwm.edu!linac!att!att!fang!gator!miles!rms
- From: rms@miles.com (Rob Schultz)
- Subject: Function points from source Code (semi-long)
- Message-ID: <1992Nov11.134743.9476@miles.com>
- Organization: Miles Inc., Diagnostics Division, Elkhart, IN
- Date: Wed, 11 Nov 1992 13:47:43 GMT
- Lines: 83
-
-
- Andrew Tune (adt@technix.oz.au) asked in private email about counting
- function points from source code. I felt that the information might be
- of general interest, so the following is the complete text of my reply
- to his query.
-
- Andrew -
-
- This is going to be fairly long, so a quick answer to you question first.
- Yes, there are tools to count function points from source code, or you
- can run a simple conversion based on the table below or in Capers Jones'
- book, _Applied_Software_Measurement_.
-
- However, there are some things to watch out for when doing this. Read on
- for more details.
-
- Software Productivity Research (Capers Jones' firm) sells its Checkpoint (TM)
- tool which allows you to estimate function points based on requirements.
- The tool also has (may be an extra option) a "backfiring" method built in.
- This allows you to take source and estimate the number of function points
- with a low degree of accuracy (+- 20% according to Capers).
-
- However, a cheaper method is to simply count LOC and use the table in his
- book to determine the size in function points. One BIG caveat with this is
- that you must use the same counting rules that Capers has used to establish
- the relationship. I have reproduced several common conversions below.
-
- This method is unsatisfactory, however. The primary reason for using
- function/feature points is to aid in the estimating process. If you are
- generating FP counts from source code, you have (hopefully) already done
- most of the work on a project (requirements, analysis, design, coding,
- test plans, etc). Therefore, you have lost the ability to predict anything
- for those phases. Of course, if you have very good data regarding the time
- it took to develop those previous versions, broken out per phase, you will
- be able to use the data to predict future projects.
-
- Capers has generated a table of phases used in construction of software and
- the time per function point to complete the phase. Look in his book for
- more information.
-
- If you would like more information, contact me (See signature), or:
-
- Software Productivity Research, Inc.
- 77 South Bedford Street
- Burlington, MA 01803
- +1 617 273 0140
-
- International Function Point Users Group (IFPUG)
- (address not handy)
- +1 614 895 1355
-
- Disclaimer: I have no affiliation with Capers Jones or SPR other than having
- just returned from two 2-day seminars on Software Measurement and Function
- Points given by Capers and Dave Herron, also of SPR.
-
- Ref:
- Jones, Capers; _Applied_Software_Measurement_; McGraw-Hill Publishing
- Company; New York; 1991; 684 pages.
-
- Jones, Capers; _Table_of_Programming_Languages_and_Levels_; Version 5.0;
- Software Productivity Research, Inc.; Burlington, MA; September
- 20, 1991; 23 pages.
-
- Language Level Source lines per FP
-
- Basic Assembler 1 320
- Macro Assembler 1.5 213
- C 2.5 128
- Cobol 3 105
- FORTRAN 3 105
- Pascal 3.5 91
- ADA 4.5 71
- Prolog 5 64
- Lisp 5 64
- Data Base 8 40
- Objective-C 12 27
- Query Languages 25 13
- Spreadsheets 30 6
-
- --
- Rob Schultz At Home: At work: +1 219 262 7206
- rms@andria.miles.com rms@miles.com
- {uunet|iuvax}!gator!miles!andria!rms {uunet|iuvax}!gator!miles!rms
-